AI That Moves the Needle: What Project Managers Need to Tell Their Executives
- Joel
- May 25
- 6 min read
Most organizations think they are watching AI from a safe distance. They are not.
AI is partially activated inside your organization's systems whether you chose it or not. The question is no longer whether you are using AI. The question is whether that use is intentional and managed, or accidental and inconsistent.
That distinction will matter more than most leadership teams currently realize.
The Pattern We've Seen Before
When a major technology arrives, organizations almost always misread it at first.
It looks like a feature upgrade — a faster way to do the same thing. A pilot runs, the tool works, and leadership feels like they are managing the transition responsibly. But in hindsight, the real change was never in the task. It was in the system. And by the time most organizations recognized that, the window to lead had already closed.
AI is following the same pattern — in two distinct phases.
Phase 1 is already underway. Emails draft faster. Reports get summarized. Routine tasks get trimmed. It looks incremental. It fits neatly into the existing operating model.
Phase 2 is what's coming: time-to-decision collapses, coordination layers shrink, project throughput rises, and firms that have built deliberate AI capability into their delivery model will complete more projects with the same team in shorter cycles. That is a structural competitive advantage — and it will change how construction services are priced and delivered.
The gap between firms experimenting and firms executing is compounding every quarter. Pilots do not close that gap. They delay the decision.
The PM Is the Right Place to Start
Before any technology conversation, it helps to be honest about what a project manager actually does.
A PM is not primarily a planner. A PM is an information hub and decision engine. On any given day, information flows in from a dozen directions — subcontractor updates, supplier delays, RFIs, change orders, cost questions, field issues, client requests. The PM synthesizes it, makes a judgment call, and communicates back out. Then does it again.
Nothing moves until the PM moves it.
That is not a criticism — it is the structural reality of how projects are organized. The PM is the integration point for an impossibly complex flow of information. And here is what that means for AI: AI is not coming for the PM's judgment. It is coming for the volume of information the PM has to manually process before they can apply that judgment.
According to Procore's 2025 research, 18% of project time is lost searching for data, and 28% is wasted due to rework. Autodesk puts the data-search problem at close to 20% of a construction professional's working hours. That is nearly half of all project time consumed by information management — before a single AI tool enters the picture.
AI closes that gap automatically. The PM arrives in the morning with a structured status brief already prepared — anomalies flagged, outstanding decisions surfaced, draft communications ready. The PM reviews, applies judgment, approves, and moves. The project moves faster because the PM moves faster.
Why Most AI Business Cases Fail
Executives hear a lot of AI pitches. Most of them fail for one of three reasons.
Tool-led cases start with "we want to implement [specific software]." That puts the focus on technology cost and triggers the wrong question: what does this tool cost? The right question is: what does our current bottleneck cost us per month?
Efficiency without dollars sounds like "we'll save four hours a week per PM." That is not a business case. Connect it to economics: faster cycle time means faster revenue recognition. Less rework means direct margin recovery. Faster decisions mean fewer delay claims.
Pilots with no scaling logic prove the technology works but never answer what it would look like embedded in how the organization actually operates. That creates false confidence, not a path forward.
The winning case for AI is operational and economic — not technical. Lead with outcomes and dollars.
A Simple Tool to Find Your Starting Point
One of the most practical frameworks for building that case is the Five Whys — a root cause analysis technique that takes five minutes and requires no training.
Pick something that slowed down a project in the last 90 days. A delay, a cost overrun, a decision that took too long. Ask why five times — each answer becoming the input for the next question.
Here's an example from a project where we applied the framework: a concrete pour was delayed. The supplier delivery was late. The order was placed later than planned. Procurement sign-off took longer than expected. The root cause: the PM didn't have the updated cost approval from finance when the order needed to go.
Run this on five recent project issues. The root cause will almost always point to one of three things: an information delay, a synthesis gap, or a communication loop. All three are squarely where AI creates leverage.
That is not a technology pitch. That is an operational diagnosis. And it lands very differently in a boardroom.
What Results Look Like in Practice
Applying the AI Adoption Framework on a $50M project, we diagnosed the highest-friction points in the delivery model and designed AI-enabled workflows around them — specifically targeting the imbalance between time spent on administration versus active risk mitigation. The results:
20–40% of PM capacity freed through improved documentation workflows — contract reviews, meeting minutes, progress tracking, monthly reports
$1M in avoided field resource costs — equivalent to 2,000–4,000 hours
Faster, more consistent communications across the project team
McKinsey's data supports the broader pattern: construction firms using AI and automation are achieving up to 20% reductions in project costs and up to 30% earlier project delivery. On one data center project, a schedule reduction of 40% was achieved using AI-driven scheduling.
The technology works. The problem, as McKinsey also notes, is that 94% of organizations that have deployed AI report they have not seen significant value from it. Nearly all of them are in Phase 1 — running pilots, enabling copilots, celebrating speed on narrow tasks — while the deeper structural changes that generate real returns remain untouched.
The Adoption Problem Is a Leadership Problem
When AI initiatives stall, the post-mortem almost always points to the same forces: fear of job insecurity, lack of trust in systems the team doesn't understand, and incentive structures that still reward individual heroics over AI-assisted performance.
None of those respond to training sessions. They respond to leadership.
Adoption is not an enablement problem. It is a change management program — with all the complexity that entails: communication strategy, leadership alignment, behavioral reinforcement, and measurement systems. Organizations consistently plan for the technology deployment and underfund the change management. That is why adoption stalls.
What actually works:
Start with augmentation, not automation. Help PMs do what they already do, better and faster. Don't begin by removing humans from decision loops.
Make AI part of the operating rhythm. If usage is optional or ad hoc, adoption will be inconsistent. Build it into standard process.
Measure behavior, not excitement. Track whether PMs are actually using the workflow and whether decisions are moving faster. Enthusiasm is not adoption.
What the Leading Firms Will Look Like
In five years, the firms at the front won't just be using AI tools. They will have redesigned their operations around AI capability.
Smaller, higher-output project teams. More projects completed with the same people in shorter cycles. Higher capital efficiency. And when throughput rises significantly, time-based billing assumptions start to shift — which changes the economics of the entire service model.
The PM who learns to work with AI is not replaced. They become the person who can manage what previously required a PM and two coordinators. That is a career advantage — and an organizational one.
The gap between the firms shaping this transition and the firms reacting to it is opening right now. Pilots won't close it. Intention will.
Joel Thompson, CPA, MBA is a principal at Redline Management Services, where he works with capital project teams on operational performance and AI adoption strategy. This post is adapted from his session at the PMI Vancouver Island Chapter Conference.
Redline Management Services works with project-driven organizations to identify where AI creates real operational leverage — and to build the workflows, governance, and adoption strategy to capture it. If you're ready to move from pilot to performance, get in touch.
